Text reading from an image and image compare plugin

[trap16]

Hello,

I'm still a noob at auto-it, and currently working on a script to ease play into a mmo (automatic crafting, travel ...), that's a great project as a first try to use autoit

I'm currently face issues which have same root, I need to identify pieces of images. Here are 2 example :

- Space orientation : There is a red arrow turning around a compass ... I'm currently detect "North" by turning slowly, until a pixel I test reach the red color , at that time I stop turning ...

That's not perfect but it works : For instance, this method will fail if unfortunately background have same color I test. And you have to find the best efficient turning speed (too fast and you skip the good pixel, too slow and it takes time to re-orientate)

- Target identification : There are lots of NPC, mobs ... once targetted a picture appear on the top of the screen, with an image of the NPC (a "living" picture unfortunately, so impossible to pixel-test) and the name of the target (fixed color for the name, but changing for the background color ...)

I have a workaround which consist in gathering strategic (part of the name, which have a constant color) coordinate to check for each NPC and try to find the correct NPC by testing them ... complicated and worse : time process consuming.

Those problems could be easily solved if a plugin or a program, or anything else could extract a text from an image (for the npc issue) and/or compare images (to be more precise, compare 2 images with a possible mask) and could interface with autoit.

I was thinking of doing myself this functionnality by getting all the pixel information and processing them ... but it's really an ambitious task ... probably much more time consuming than the script I'm working on actually, so I'm asking you have any idea which can help me in finding a solution ?

Many thanks.

Olivier